knitr::opts_chunk$set(echo = TRUE, message = FALSE, warning = FALSE)

library(tidyverse)
library(here)
library(sf) # install.packages('sf')
library(tmap)
### install.packages('tmap')
### update.packages(ask = FALSE)

cmd-shift-enter shortcut for running the current code chunk

Read in the data

cmd-option-i shortcut for creating a code chunk

sf_trees <- read_csv(here('data', 'sf_trees', 'sf_trees.csv'),
                     show_col_types = FALSE)

names(sf_trees) to get the names of the columns summary(sf_trees) to get an idea about the data in each of those columns

Part 1: wrangling and ggplot review

Example 1: Find counts of observation by legal_status & wrangle a bit.

### method 1: group_by() %>% summarize()
sf_trees %>% 
  group_by(legal_status) %>% 
  summarize(tree_count = n())
## # A tibble: 10 × 2
##    legal_status                 tree_count
##    <chr>                             <int>
##  1 DPW Maintained                   141725
##  2 Landmark tree                        42
##  3 Permitted Site                    39732
##  4 Planning Code 138.1 required        971
##  5 Private                             163
##  6 Property Tree                       316
##  7 Section 143                         230
##  8 Significant Tree                   1648
##  9 Undocumented                       8106
## 10 <NA>                                 54
### method 2: different way plus a few new functions
### store stuff in an object this time
### highlight code and press command+enter
top_5_status <- sf_trees %>% 
  count(legal_status) %>% 
  drop_na(legal_status) %>% 
  rename(tree_count = n) %>% 
  relocate(tree_count) %>% 
  slice_max(tree_count, n = 5) %>% 
  arrange(desc(tree_count))
### drop_na() would get rid of a row with NA in any column. we just want NA in legal_status column not for example na in a date column
### rename column from n to tree_count
### relocates that column to the front of the dataframe
### look at all values of tree_count and slice out the highest ones. in this case the top 5
### arrange(tree_count) does lowest to highest
### to get highest to lowest: arrange(-tree_count) OR arrange(desc(tree_count))

Make a graph of the top 5 from above:

ggplot(data = top_5_status, aes(x = fct_reorder(legal_status, tree_count), y = tree_count)) +
  geom_col(fill = 'darkgreen') +
  labs(x = 'Legal status', y = 'Tree count') +
  coord_flip() +
  theme_minimal()

### to change the order of the columns: fct_reorder(legal_status, tree_count) to order the trees by tree_count. if wanted largest to smallest, then do -tree_count

Example 2: Only going to keep observations where legal status is “Permitted Site” and caretaker is “MTA”, and store as permitted_data_df

shift-cmd-c to comment/uncomment quickly

# sf_trees$legal_status %>% unique()
# unique(sf_trees$caretaker)
permitted_data_df <- sf_trees %>% 
  filter(legal_status == 'Permitted Site', caretaker == 'MTA')
### can also use & instead of ,
### use | as or
### filter(legal_status %in% c('Permitted Site', 'Private') & caretaker == 'MTA')

Example 3: Only keep Blackwood Acacia trees, and then only keep columns legal_status, date, latitude, longitude and store as blackwood_acacia_df

blackwood_acacia_df <- sf_trees %>% 
  filter(str_detect(species, 'Blackwood Acacia')) %>% 
  select(legal_status, date, lat = latitude, lon = longitude)
### rename column to lat

### Make a little graph of locations
ggplot(data = blackwood_acacia_df, aes(x = lon, y = lat)) +
  geom_point(color = 'darkgreen')

### doesn't know this is spatial data yet

Example 4: use tidyr:separate()

sf_trees_sep <- sf_trees %>% 
  separate(species, into = c('spp_scientific', 'spp_common'), sep = ' :: ')

Example 5: use tidyr::unite()

ex_5 <- sf_trees %>% 
  unite('id_status', tree_id, legal_status, sep = '_COOL_')

Part 2: make some maps

Step 1: convert the lat/lon to spatial point, st_as_sf()

### sf is simple features object
### it knows this is a spacial object with st_as_sf
blackwood_acacia_sf <- blackwood_acacia_df %>% 
  drop_na(lat, lon) %>% 
  st_as_sf(coords = c('lon', 'lat'))
  
### we need to tell R what the coordinate reference system is
st_crs(blackwood_acacia_sf) <- 4326

ggplot(data = blackwood_acacia_sf) +
  geom_sf(color = 'darkgreen') +
  theme_minimal()

Read in the SF shapefile and add to map

sf_map <- read_sf(here('data', 'sf_map', 'tl_2017_06075_roads.shp'))
### read in the shapefile

### st_crs(sf_map)

sf_map_transform <- st_transform(sf_map, 4326)

ggplot(data = sf_map_transform) +
  geom_sf()

Combine the maps!

ggplot() +
  geom_sf(data = sf_map, #layers build. this layer will be underneath
          size = .1, # make size of lines smaller
          color = 'darkgrey') +
  geom_sf(data = blackwood_acacia_sf,
          color = 'red',
          size = 0.5) +
  theme_void() +
  labs(title = 'Blackwood acacias in SF')

Now an interactive map!

tmap_mode('view')

tm_shape(blackwood_acacia_sf) +
  tm_dots()